Cognitive Reserve in Granulin-Related Frontotemporal Dementia: from Preclinical to Clinical Stages Enrico Premi1, Stefano Gazzina1, Marco Bozzali2, Silvana Archetti3, Antonella Alberici1, Mara Cercignani4, Angelo Bianchetti5, Roberto Gasparotti6, Marinella Turla7, Carlo Caltagirone8, Alessandro Padovani1, Barbara Borroni1* 1 Centre for Neurodegenerative Disorders, University of Brescia, Brescia, Italy, 2 Neuroimaging Laboratory, Santa Lucia Foundation IRCCS, Rome, Italy, 3 III Laboratory of Analysis, Brescia Hospital, Brescia, Italy, 4 Brighton and Sussex Medical School, Clinical Imaging Centre, University of Sussex, Brighton, United Kingdom, 5 Geriatric Research Group, Brescia, Italy, 6 Neuroradiology Unit, University of Brescia, Brescia, Italy, 7 Neurology Unit, ValleCamonica Hospital, Brescia, Italy, 8 Department of Neuroscience, University of Rome “Tor Vergata”, Rome, Italy
Abstract Objective: Consistent with the cognitive reserve hypothesis, higher education and occupation attainments may help persons with neurodegenerative dementias to better withstand neuropathology before developing cognitive impairment. We tested here the cognitive reserve hypothesis in patients with frontotemporal dementia (FTD), with or without pathogenetic granulin mutations (GRN+ and GRN-), and in presymptomatic GRN mutation carriers (aGRN+). Methods: Education and occupation attainments were assessed and combined to define Reserve Index (RI) in 32 FTD patients, i.e. 12 GRN+ and 20 GRN-, and in 17 aGRN+. Changes in functional connectivity were estimated by resting state fMRI, focusing on the salience network (SN), executive network (EN) and bilateral frontoparietal networks (FPNs). Cognitive status was measured by FTD-modified Clinical Dementia Rating Scale. Results: In FTD patients higher level of premorbid cognitive reserve was associated with reduced connectivity within the SN and the EN. EN was more involved in FTD patients without GRN mutations, while SN was more affected in GRN pathology. In aGRN+, cognitive reserve was associated with reduced SN. Conclusions: This study suggests that cognitive reserve modulates functional connectivity in patients with FTD, even in monogenic disease. In GRN inherited FTD, cognitive reserve mechanisms operate even in presymptomatic to clinical stages. Citation: Premi E, Gazzina S, Bozzali M, Archetti S, Alberici A, et al. (2013) Cognitive Reserve in Granulin-Related Frontotemporal Dementia: from Preclinical to Clinical Stages. PLoS ONE 8(9): e74762. doi:10.1371/journal.pone.0074762 Editor: Stefano L Sensi, University G. D'Annunzio, Italy Received June 15, 2013; Accepted August 2, 2013; Published September 9, 2013 Copyright: © 2013 Premi et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The authors have no funding or support to report. Competing interests: The authors have declared that no competing interests exist. * E-mail:
[email protected]
Introduction
require more brain atrophy in those regions typically targeted by the pathology to exhibit the same level of cognitive decline shown by AD patients with lower education levels [3]. In the same view, the role of cognitive reserve hypothesis has been investigated also in Frontotemporal Dementia (FTD), a neurodegenerative disease characterized by behavioral disorders, language impairment, and deficits of executive functions as most typical clinical features [4,5]. Literature data suggested that education and occupational attainments might act as a proxy measure of reserve capacity in FTD, as well as AD [6]. Furthermore, as in AD [7], cognitive reserve in FTD is still in action even in the presence of an unfavorable genetic background [8]. FTD has a strong genetic background, and a number of genes causative of autosomal dominant forms have been
The cognitive reserve hypothesis posits that lifetime intellectual enrichment lessens the negative impact of neurological diseases on the cognitive status [1]. When the neurocognitive processing is challenged by brain disease, individuals with greater premorbid cerebral efficiency are able to withstand better the neurocognitive challenges, thus showing a relative resilience to cognitive impairment [1]. To account for these clinical observations, the concepts of cognitive and brain reserves have been developed [2], with the hypothesis that phenomena of brain plasticity might represent the underlying neurobiological substrate. It has been recently demonstrated that Alzheimer’s disease (AD) patients with higher levels of formal education (a proxy measure of cognitive reserve)
PLOS ONE | www.plosone.org
1
September 2013 | Volume 8 | Issue 9 | e74762
Cognitive Reserve in Frontotemporal Dementia
identified so far. Among others, Granulin (GRN) mutations, inducing a loss of 50% functional Progranulin [9,10], are present in a proportion of patients whose most typical clinical presentations include the behavioral variant of Frontotemporal Dementia (bvFTD) and the agrammatic variant of Primary Progressive Aphasia (avPPA). GRN mutations are, by definition, inherited at birth, with the disease onset that typically occurs at the 5th-6th decade of life, although there are rare subjects who carry pathogenetic variation in their late life, without any sign of the disease. This means that FTD patients carrying GRN mutation have a completely normal life until their fifties, and if the disease begins, GRN mutation carriers have a worse clinical prognosis than FTD patients without mutations [11]. However, a small quote of mutation carriers show an incomplete penetrance, thus suggesting the possibility of genetic or environmental disease modifiers. As many cases of FTD are inherited, the role of cognitive reserve in patients with monogenic disease, i.e. GRN-disease, still needs to be established moving from preclinical to symptomatic stages. Imaging genetics is a growing field that is shedding light for new discoveries in neuroscience [12]. Magnetic resonance imaging (MRI) has become an increasingly powerful tool for human brain investigation, and using different modalities, has been successfully used to investigate different pathophysiological aspects of the brain tissue in the presence of neurodegeneration [13,14]. Beyond structural MRI, resting state functional MRI (fMRI) has shown the ability to provide measures of functional brain connectivity, based on the evidence that different brain regions are functionally synchronized at rest, and connected regions are supposed to define common networks subserving complex brain functions. In the presence of neurodegeneration, the loss of functional brain connectivity is likely to account for cognitive disabilities and even for some gray matter loss secondary to neuronal disconnection [15]. From resting state fMRI data (i.e., fMRI time series collected while subjects lie vigilant but at rest in the scanner), several networks can be extracted in a data-drive fashion, by using the so-called Independent Component Analysis algorithm [16]. Initial resting-state studies in FTD described a divergent relationship between Default Mode Network (DMN) and Salience Network (SN) connectivity, with attenuated connectivity of SN [17,18], whose activity is related to the autonomic/interior processing and the "salience" of the stimulus, like the emphatic mechanisms and the emotional aspect of pain [19,20]. However, recently other networks have been described as involved in FTD, in particular the Executive Network (EN), and Frontoparietal Networks (FPNs) [21]. The areas belonging to EN have been hypothesized to provide bias signals to other areas of the brain in order to improve cognitive control [22]. Furthermore, the cortical regions sustaining EN are specifically involved in Frontotemporal Dementia, playing a role in the disease progression [23]. On the other side, FPNs have been related to top-down modulation of attention and working memory [24]. From previous studies FPNs seem to be involved in the selection of relevant environmental information, which could be important for the integration between environmental sensory stimulus and behavioral goals and
PLOS ONE | www.plosone.org
expectations [25]. Furthermore, in FTD these changes are more pronounced in patients with GRN mutations; at the moment, only two studies have explored functional network connectivity alterations in presymptomatic GRN carriers showing impaired resting state functional connectivity in the network primarily involved in the pathology (i.e. SN) [18,26]. Resting state fMRI may contribute to clarify the interaction between genetic and environmental factors in modulating the occurrence of clinical symptoms and to define a theoretical model of disease progression, moving from the presymptomatic stage to clinical presentation. With these caveats in mind, the current study uses resting state fMRI to investigate the relationship between lifetime intellectual enrichment and patterns of brain connectivity in patients with FTD, with and without GRN pathogenetic mutations, and in presymptomatic GRN mutation carriers.
Methods Subjects Subjects entering the present study were partly the same as those recruited for a previous investigation [18] (N= 30), and in part (N=19) newly recruited. In the former case, subjects were invited to attend again the Centre for Ageing Brain and Neurodegenerative Disorders, at University of Brescia (Brescia, Italy), to collect data for the assessment of cognitive reserve. In the latter case, subjects were also asked to undergo the MRI protocol, as detailed below. The studied sample included 32 patients with FTD all genetically characterized for the presence/absence of GRN and MAPT mutations and C9orf72 hesanucleotide expansion. Twelve of them proved to be carriers of GRN Thr272fs mutation (GRN+), while the remaining 20 proved to be non-carriers of screened genetic variations (GRN-). The current study included also 17 asymptomatic carriers of GRN Thr272fs mutation (aGRN+; all siblings of GRN+ FTD patients). Nine of them had already taken part in our previous study [18], while the remaining 8 were newly recruited. All FTD patients met current clinical diagnostic criteria for bvFTD [27] (18 cases) or avPPA [28] (14 cases). To increase as much as possible the confidence of a correct diagnosis of FTD in patients without GRN Thr272fs mutation, they had to be clinically and neuropsychologically followed-up for at least 2 years, at the time of recruitment. All patients underwent a clinical and neurological evaluation, a routine laboratory examination, and conventional brain MRI before entering this study, to rule out any potential alternative diagnosis. An extensive neuropsychological assessment in both patients and asymptomatic siblings, including the FTDmodified Clinical dementia Rating scale (FTD-modified CDR), was administered, as previously described [18]. Written informed consent (from the subject or from the responsible guardian if the subject was incapable) was obtained, for each procedure, before study initiation, including blood collection from venous puncture, genetic analysis, and MRI scanning. The research protocol was approved by the ethics committee of the Hospital (Comitato Etico, Azienda
2
September 2013 | Volume 8 | Issue 9 | e74762
Cognitive Reserve in Frontotemporal Dementia
were filtered by a phase-insensitive bandpass filter (pass band 0.01–0.08 Hz) to reduce the effect of low frequency drift and high frequency physiological noise. Briefly, group ICA for fMRI toolbox first concatenates the individual data across time, and then produces a computation of subject specific components and time courses. For all subjects grouped together, the toolbox performed the analysis in 3 steps: (1) data reduction, (2) application of the FastICA algorithm, and (3) back-reconstruction for each individual subject [30]. ICA analysis was employed to identify 40 independent components, using the Minimum Description Length Criterion for the dimension determination [31]. Statistical reliability of independent component decomposition was evaluated using the ICASSO Toolbox, implemented in GIFT [32] running FastICA algorithm 10 times with different initial conditions and bootstrapped data sets. Results were converted to Z-scores. The 40 components were reviewed, and compared, by computing the spatial correlation coefficient, to customized templates of the networks affected by the pathology, according to literature data [21] i.e. dorsal and ventral Salience Network (SN), Default Mode Network (DMN), Executive Network (EN), Frontoparietal Networks (FPNs) and Dorsal Attention Network (AN) [21]. This procedure was performed using the tool for spatial sorting of the components available with GIFT. Every subject’s Z-score maps corresponding to these resting state networks were used for cross-subject analyses. For the purpose of the present study, subjects were divided into 3 separate groups: patients with FTD GRN Thr272fs mutation carriers (GRN+, n = 12); patients with FTD non mutation carriers (GRN-, n = 20); asymptomatic subjects FTLD GRN Thr272fs mutation carriers (aGRN+, n = 17). Age, gender, dementia severity scored with FTD-CDR scale and total grey matter volume were entered as covariates of no interest. For each considered network, contrasts were designed to assess the correlation of RI with functional connectivity in FTD patients (either in FTD-GRN+ and FTD-GRN-); at this purpose, a linear regression analysis between RI (as independent variable) and network resting-state functional connectivity (dependent variable) was performed in each group (FTD-GRN+ and FTD-GRN-); then, a difference of slope (interaction analysis) was used to evaluate different reserve effects in GRN + and GRN-; in particular, the statistical differences between the regression of RI scores in FTD-GRN+ and FTD-GRNgroups were studied (FTD-GRN+ < FTD-GRN-; FTD-GRN- < FTD-GRN+). [33]. In the aGRN+ group a linear regression analysis using RI scores was performed to study the reserve effect. P-values were defined at p
